Installing bird routing package

[mazzy89]

I would need to install bird on the OS to distribute route of pods but I could not find a good way to do that w/o a package manager. My last resort would be to run it on a container and give privileges to access the kernel and the nework of the machine. what are the options here?

[mazzy89]

I was thinking to add bird here https://github.com/rancher/k3os/blob/master/images/00-base/Dockerfile. What do you think? Would be fine to make an ISO and make upstream with that dependency?

[dweomer]

I was thinking to add bird here https://github.com/rancher/k3os/blob/master/images/00-base/Dockerfile. What do you think?

If you are fine with the version available from the Alpine base, that is a great place to add it.

Would be fine to make an ISO and make upstream with that dependency?

You are always free to build your own k3OS ISO. I do not understand "and make upstream with that dependency" however. If you are asking would we accept a PR for such a change? I would consider it but I must confess that new releases are unlikely at this point.

[mazzy89]

Why do you say that new releases are unlikely? Is the project discontinued?

[mazzy89]

@dweomer please could you shed some light on the statement

I would consider it but I must confess that new releases are unlikely at this point.

Looking at the releases, the last was In October. k8s latest release at the time of writing is 1.23 but 1.24 is going to be very soon after the Christmas holidays.

That being said could you clarify the future of this project?

[mazzy89]

trying to break the radio silence here and understand the future of this project for this year.

@dweomer could you shed some light on it? Should we as users get ready to abandon the ship and look for greener pastures?

Considering that now 1.24 is out and the latest stable release is 1.21, are then any releases to the horizon? is the project going to be discontinued for some internal reasons at rancher?

[themgt]

My vague impression is when SUSE bought Rancher some sort of decision was made to deprioritize / reevaluate the k3os (but not k3s) project. The exact reason was never made clear, but it seems like SUSE owes the community some sort of clarity on what's going on. I don't think this is a burned out OSS maintainer story but a corporate decision.

It's plausible that the community could take over k3os maintenance from SUSE if needed, but the lack of communication/direction is making that difficult.

[thejimnicholson]

It would be good for SUSE to have a little more transparency about the future of k3os. I've found it useful (and agree with the philosophy behind it,) and I'd be happy to add whatever I could to an effort to maintain it (with the proviso that I don't actually know what I'd be doing in that regard. I've got quite a bit of personal time and effort invested in the 2 k3os clusters I'm running; they essentially are the basis for all my k8s knowledge at this point. I'd like to see the project continue, but if Rancher/SUSE isn't prepared to support that effort, perhaps it's time for a fork?

[j-landru]

A contributor of kube-hetzner post a k3s issue k3s-io/k3s#4991 to k3s team requesting to "move k3os under k3s-io's wing"...
Hoping that it will receive a positive consideration.